An Overview of HIP Research Topics

Broad and complementary expertise of the partners

We at HIP understand Imaging as a pipeline that covers the full range of data acquisition, data preparation, data management up to data analysis. Every part of this pipeline comes with its specific scientific challenges and research software tools. The HIP core team partners focus on different parts of the pipeline corresponding to their main expertise. This joint approach across communities and research fields provides a unique opportunity to combine skills and strength of single centres for the benefit of the entire Association.

All HIP core team partners are united by frontline research and extensive imaging competence, which provides an excellent starting point and competitive basis for first-class, state-of-the-art services provided by the HIP core units. For all parts of the imaging pipeline, the HIP core team will be able to systematically find and leverage synergies across modalities and research questions, thus carving out common challenges and generalizing solutions across the Helmholtz Association.

HIP Imaging Pipeline

Image data acquisition (DESY)

At the beginning of the imaging pipeline, data is acquired measuring the change of an emitted signal. This change is generated by interaction with the sample and can be measured physically on the one hand and modelled mathematically on the other. For a known sample, the response of the physical system can be determined from the model. Far more often, however, one would like to infer the nature of the sample from the measured response. To do this, the mathematical model must be inverted. These so-called inverse problems are at the heart of almost any image formation process.

DESY has a long-standing experience in solving inverse problems, such as the phase problem in crystallography and coherent diffraction imaging and as the tomographic problem in many different variants. The HIP scientific unit at DESY will be a research group in applied mathematics and computer science, specialized in the field of inverse problems from a mathematical point of view. This group shall bundle the expertise in inverse problems in a generic way, thus being able to support developments of other groups involved in image formation across the Helmholtz Association.

Image data preparation & management (MDC)

The MDC will place special emphasis on the integration of image data, algorithms and visualization solutions across multiple modalities and scales in space and time. The goal is to develop and provide HIP solutions that can handle the very heterogeneous image data collected across the research fields of the Helmholtz Association without imposing restrictions on the respective image modalities. In order to lay the foundations for the implementation of HIP solutions, the MDC will focus on the following research topics:

  1. Develop concepts and algorithms for handling and generic processing of high-dimensional datasets
  2. Develop algorithms for large, high-dimensional image data stitching, fusion and visualization

Image data analysis (DKFZ)

The scientific unit at DKFZ will address important bottlenecks in Helmholtz imaging research concerning the annotation and analysis of imaging data:

  1. The manual and semi-automatic labelling of large amounts of imaging data - a core prerequisite for performing AI-based imaging research. In-depth competencies in this area yield a large potential for overarching synergies within Helmholtz imaging sciences.
  2. The automated analysis and information extraction from imaging, including semantic segmentation, detection of change, novelty of objects, and the estimation of pose and object tracking in time-resolved image series.
  3. The facilitation and implementation of open benchmark competitions for the validation of findings to ensure scalability, reproducibility and applicability of the imaging methods developed in Helmholtz.